How Did We Get Here?

What happened to all the children in San Francisco?

Thanks to the hundreds of new subscribers who signed up when school closures were under discussion. I hope you continue to find this an interesting read even though that topic has been pushed to the back burner.

We have a new school board and a new superintendent. While they get to grips with their portfolios, we have time to take a step back and consider how San Francisco has evolved to be a city where children are scarce.

1950

Let’s start all the way back in 1950. California was booming. The state’s population had risen by over 50% in the 1940s and the baby boom was under way. Much of the growth was outside the main cities. Contra Costa’s population had tripled in the 1940s. San Mateo’s had more than doubled. Richmond had grown by 320% in the decade to 100,000 people and was the 9th largest city in the state. Oakland was the 3rd largest city, bigger than San Diego. Berkeley (7th) and Richmond (9th) were both bigger than San Jose (10th)¹.

San Francisco grew by only 22% but those additional 140,000 people brought the population to 775,357, a number it wouldn’t surpass for 50 years. Demographically, the city was very different from today. 89% of the population was White (but “White” included Latino). There were 101 males for every 100 females, a completely unexceptional ratio. The household median income of $3,005 was just below the state median of $3,021. Educational achievement, as measured by median school years completed, was average.

But the city in 1950 already had some characteristics that it retains to this day.

• The median age was 36.2, well above the state’s median age of 32.1. In the other large Bay Area cities, the median age was 31.4 in Berkeley, 34.6 in Oakland, 25.9 in Richmond, and 32.7 in San Jose.

• The average household had 2.7 persons, well below the state average of 3.01. No county had a smaller average household size but some cities did: Santa Cruz was at 2.50, Long Beach at 2.65, Modesto at 2.66.

• People were less likely to be married in San Francisco² than in other cities.

Over a quarter of the 140,000 newcomers to the city were Black and their arrival would have been very visible because it increased the Black population nearly nine-fold to 44,000 and meant they outnumbered the city’s longstanding Asian population.“Asian” wasn’t even a category in the census. The main census categories were Native White, Foreign-born White, Negro, and Other Races. Immigration restrictions also meant that foreign-born whites had a median age of 54 while native whites had a median age of 32.2. We have to go to a subsidiary table to learn that there were nearly 25,000 Chinese in the city, albeit with a very skewed gender ratio of 169 males per 100 females, a visible legacy of the restrictions on who could immigrate. The Japanese population was smaller (5,600) but had a more normal ratio of 96 males per 100 females.

At that point, there were 159,000 people under 18 in the city. Thanks to the baby boom 63,000 were under 5, so the school-age population was about 96,000, of which SFUSD enrolled 66,000. We don’t know what fraction of the remainder were in private school or not in any school. Some would have been in kindergarten which did not count as being “in school” for census purposes back then.

1950-1960

The 1950s were the heart of the baby boom and California was still a magnet for domestic migration. The state’s population rose by over 5,000,000 (48%), still the second biggest increase in any state’s population in any decade. Imagine all the construction of both houses and schools that was needed to accommodate those new arrivals. Again, growth was highest outside the major cities where there was room. Santa Clara grew by 121% as Silicon Valley took root. San Mateo grew by 88%. Within San Mateo, Daly City grew by 195% and San Bruno by 133%. Alameda grew 23% and became the first Bay Area county to surpass San Francisco in population.

San Francisco did not share in the growth. Its population fell by 35,000 (5%) because the phenomenon of White flight was underway. Cities across the country that had seen an influx of Black people saw an exodus of White people. Within California, the only other places to see population declines were Oakland, Albany, Berkeley, and Richmond. During the 1950s, the White population of San Francisco fell by 90,000, many to those growing towns on the peninsula, and they were only partially replaced by new Black and Asian arrivals.

The population per household fell to 2.44. The only places lower were those with few kids and lots of old people like Beverly Hills and Carmel-by-the-sea. The gender ratio stood at 96.7 men aged 25-64 for every 100 women. The ratio for Chinese (of all ages) had improved to 130 and Filipino residents now had the most skewed gender ratio: 153 males per 100 females.

The baby boom was ending but, by 1960, the number of people under 18 reached 181,000. This would be its peak. Nearly 25% of the city’s population was under 18 but that was still far below the rates in some of the neighboring cities³. 47% of Pacifica’s population was under 18. That’s a sub-Saharan Africa rate but other cities were not far beyond. The fraction under 18 was 41% in South San Francisco, 39% in San Bruno, and 38% in Daly City.

SFUSD had 88,000 students, meaning it had added 22,000 students in a decade. Imagine the challenges that posed. For the first time, we have data on private school enrollment: 25% of elementary school and 22% of high school students were in private schools. Kindergarten was still considered to a separate thing from elementary school but 90% of five and six year olds were now enrolled in school.

1960-1970

The 1960s saw a continuation of the trends of the 1950s. California added another 4,000,000 (27%) people while growth in the Bay Area ranged from 18% in Alameda to 66% in Santa Clara. That growth made Santa Clara the second Bay Area county to surpass San Francisco’s population.

Meanwhile, San Francisco lost another 25,000. There were now only 2.34 persons per household. Another 90,000 White residents left the city and the Asian and Black arrivals were not sufficient to completely replace them. These two groups were now roughly the same size. There were 96,000 Black and 95,000 Asian residents and they both outnumbered the 76,000 Latino⁴ residents. Although no-one knew it at the time the Black population had now reached its peak.

The gender ratio in the 25-64 group was 96.7 men for every 100 women but, for people in their 30s, the ratio was over 110 men per 100 women. The city was acquiring its reputation as a haven for gay men.

The end of the baby boom and the ongoing exodus of families meant that the number of residents under 18 fell by 20,000 to 160,000. Most of that decline took place in the under 5 population. Because it took time for the baby boomers to make their way through the school system, the number of students in public school was still 89,000, a bit below its peak of 92,150 in 1963-64 but above where it had been in 1960. Private school enrollment did fall, by 13%, due to all the White families leaving the city.

The end of the baby boom had a similar effect throughout the Bay Area. Public school enrollment peaked in 1968-69 in Contra Costa, Marin, and San Mateo. It peaked in Alameda in 1970-71 and in Santa Clara in 1972-73. Subsequent population growth would eventually enable Contra Costa, Alameda, and Santa Clara to break those records but San Francisco, Marin, and San Mateo have never matched them.

1970-1980

The 1970s saw the culmination of the post-war trends. California added another 3.7 million (19%). Growth in the Bay Area ranged from 3% in Alameda to 22% in Santa Clara. San Francisco lost another 37,000 (5%) driven by the loss of another 116,000 White residents. The decline in the 1970s was greater than that in the 50s or 60s for two reasons. In the 1970 census, 96% of “persons of Spanish language or Spanish surname” were classified as White. By 1980, that number was down to 49% with nearly 35,000 moving into what is today called the Other category. The other new impetus for the decline was SFUSD’s introduction of busing to integrate schools. Ironically, just as the district was attempting to enact policies specifically to benefit them, the Black population started its long and ongoing decline. Replacing the departed were Asian and Latino arrivals but the city’s overall population still reached its post-war nadir of just under 680,000 in 1980.

The number of persons per household dropped to 2.18. In the 25-64 age range, there were now 107.5 men for every women. Among those aged 35-39, the ratio was 122.9. This dramatic change was obviously driven by the influx of gay men during the 1970s.

The last of the baby boomers graduated high school during the 1970s so the number of residents under 18 fell by 28%, much more than the overall population decline, to 115,000.

Public school enrollment cratered, going from 89,000 at the start of the decade to 56,000 at the end of it. Private school enrollment didn’t fall. It even ticked up by a few hundred students. It’s easy to imagine families who didn’t like the idea of busing enrolling their children in private schools as an alternative to moving out of the city. Busing had ended by 1980 but perhaps those families were reluctant to move back. Private schools now taught 33% of elementary school students and 25% of high school students. This remains the highest market share for private schools.

1980-1990

The 1980s saw renewed growth. California’s population grew by another six million (26%), still the highest increase in any state’s population in any decade. Bay Area growth ranged from 3% in Marin to 22% in Contra Costa. Contra Costa became the third Bay Area county to surpass San Francisco.

San Francisco experienced its first growth since the 1940s, adding 45,000 (7%) new residents. The non-Latino White population fell by over 20,000 (7%) and the Black population by 7,000 (9%) but the Asian population grew by over 60,000 (43%) and the Latino population by 17,000 (21%). Household size even grew a bit, to 2.29.

Despite the AIDS epidemic, the male-female ratio increased in the 25-64 segment to 109.4 men per 100 women, although it did slip in the narrower 30-39 segment.

The under 18 population showed a tiny 1.5% increase but this was not spread evenly. A mini baby-boom pushed up size of the under 5 population but the school age population fell by 3,000 to 81,000. Despite this, SFUSD’s enrollment rose by over 6,000 (11%) to 62,500. The difference is accounted for by a big drop in private school enrollment once the threat of busing had passed.

1990-2000

The 1990s continued the pattern of the 1980s. California added another four million people (14%) and growth in the Bay Area ranged from 7% in Marin to 18% in Contra Costa.

The second-half of the 1990s was the dot com era, the first time businesses with no idea how to make money could raise a lot of capital and hire a lot of people. San Francisco became home to its first tech companies. Its population grew by 53,000 (7%) to 777,000, putting it slightly above previous high which was from 1950.

Perhaps attracted by dot com jobs, the number of non-Latino White people rose. It’s hard to say precisely how much because this was the first census that offered a Two or More option and some fraction of the 23,000 Two or More people would have classified themselves as White in previous censuses. Meanwhile, the Asian and Latino populations grew by 34,000 and 9,000 respectively and the Black population shrank by over 20% (again, some of these losses would not have been departures from the city but individuals reclassifying themselves into the Two or More category).

The south of market area saw a lot of housing development for the first time. Because of housing construction, the population per household stayed flat at 2.30. The 25-64 gender ratio moved up from 109.4 to 111.7.

Despite the population growth, the number of those under 18 dropped by 4,000 (4%). A drop in the birth rate meant that all of that decline was in the 0-5 age group. The school age population and the number in public school changed by only a few hundred. SFUSD, however, experienced a drop of over 2000 in part because over 500 students were added to the rolls of its alter ego, the San Francisco County Office of Eduction, and in part because of the introduction of charter schools.

2000-2010

This was an eventful decade. It started with the dot-com crash, followed by the expansion and collapse of the property bubble. California added another 3.4 million people but this represented only 10% growth. In the Bay Area, only Contra Costa had higher growth that the state average.

San Francisco added 29,000 new residents, putting it above 800,000 for the first time. Ongoing construction meant that the population per housing unit fell from 2.30 to 2.26. The new residents continued to be Asian and Latino. The non-Latino White population was flat and the Black population fell by another 9,000 (15%).

The gender ratio among those aged 25-64 slipped from 111.7 to 111.2. The ratio was highest among those in their 40s, but this was the first time that men outnumbered women in every five-year cohort between 25-29 and 60-64.

Having fallen in the 1990s, the number of births grew in the 2000s and this drove up the Under 5 population. Nevertheless, the school age population fell by nearly 9,000 (11%). Private schools lost 3,000 students (16%), while SFUSD lost 7,000 (11%). New and existing charter schools added 1,500 (136%).

2010-2020

California’s growth slowed this decade. It added only 2,300,000 people, the fewest number of people in any decade since the 1940s, and the lowest percentage growth (6%) ever. Curiously, all the Bay Area counties showed higher growth in the 2010s than in the 2000s, led by 11% in both Alameda and Contra Costa. Marin, as usual, brought up the rear.

San Francisco’s population grew faster than the state of California’s for the first time since the 1870s. That’s so long ago that California’s population in 1880 was smaller than San Francisco’s in 2020. San Francisco added nearly 70,000 (8.5%) new residents but new construction, most visibly in Mission Bay, meant that the population per household barely moved, from 2.26 to 2.28.

Half the newcomers were Asian (9% growth) while the Latino population grew 15,000 (12%). The White population stayed stable while the Black population slipped another 4%, leaving it pretty much at its 1950 level.

Meanwhile, the gender ratio in the 25-64 age group edged up to 111.5 men per 100 women from 111.2. The greatest disproportion, 118.2, was to be found among those aged 50-54.

The school-age population grew by about 3,500 (5%). SFUSD enrollment was down a few hundred but charters grew 2,000 (74%) and private schools grew by about 600 (2.5%).

2020-today

It is difficult to get an accurate view of how much the pandemic affected the city’s population. The census bureau’s latest population estimate for the city is about 809,000, down nearly 65,000 from the 874,000 recorded in the census. We know that school enrollment is down about 4,000 but I don’t have a lot of confidence in the Census Bureau’s estimate⁵. California’s Department of Finance has its own demography unit and its most recent estimate is 843,000, a decline of only 31,000.

The Present and Future of the Child Population

The city is tasked with adding lots of new housing units over the next decade but, as San Francisco’s own experience has shown, increasing the number of housing units might be offset by fewer children per household and not result in any increase in the number of children.

Although San Francisco’s population grew by nearly 200,000 (29%) between 1980 and 2020, the number of children under 18 fell by 2%. The fraction of the population that is under 18 has fallen every decade since 1960 and is now 13%. Some retirement destinations (e.g. The Villages in Florida) and college towns (e.g. Ithaca, New York) have even lower percentages but, when expressed as a percentage of the population aged 25-64, none of the 250 largest cities in the country has fewer kids than San Francisco.

Population Distribution

San Francisco has become a city filled with young adults who are destined to spend only a few years in the city before moving out when they get older and have families. 32.1% of San Franciscans are aged 25-39, compared to only 11.0% who are aged 0-14. It is common to think of San Francisco as being a mini-Manhattan but, in some ways, San Francisco is more extreme than Manhattan. Although Manhattan’s population density is four times that of San Francisco⁶, it is not as dependent on young adults (30.3% are in the 25-39 range) and has more children (11.5% are in the 0-14 range) than San Francisco. In fact, the rest of New York City (i.e. excluding Manhattan) is 37% more dense than San Francisco but has 64% more kids.

San Francisco’s adult (i.e. aged 25-64) gender ratio of 111.5 men per 100 women is the highest of any large city. The majority of cities are balanced like Los Angeles or skew female like Manhattan (93.5). Boston (91.1), and Chicago (96.9). San Francisco’s gender ratio may have originally skewed male because of the large gay community but, today, the tech sector is another contributing factor. Other large cities that are tech hotspots also have male-skewed ratios, including Seattle (107.0), Austin (106.8), Salt Lake City (110.8), and San Jose (103.3). Smaller tech cities in California also skew male, including Milpitas (109.0), Mountain View (112.9), Santa Clara (109.8), and Sunnyvale (110.8).⁷

Of all the data I looked at before writing this post, by far the most surprising to me was that shown in the next chart. There are far more men than women in the White, Latino, and Black communities but far more women than men in the Asian community. I had no idea that the gender ratio varied so much by race. It’s not really relevant to this post but it was so unexpected that I can’t resist sharing it.

Births

Births to San Francisco residents have been under 7,000 per year for the last year, the lowest since records began. The previous low of just under 7,400 was reached in 1975.

The birth records show significant fluctuations over the decades:

• A rapid 50% decline from the 1960s through 1975.

• Consistent growth from 1975 through to 1990. From trough to peak, births rose 37%.

• A 20% decline during the 1990s.

• A modest rise of about 12% from 2000 to 2017.

• A 26% decline in just seven years since 2017.

Explaining this pattern is not easy. Why did births decline during the 1990s or after 2017? The pandemic accelerated the decline but didn’t cause it. By February 2020, births were already down to a number last seen in 2002. The temptation is to look for local explanations but other counties saw their birth numbers fall earlier. From 2000 to 2022, births across the state were down 21%, compared to “only” 18% in San Francisco. The biggest decline (39%) was in Los Angeles but, even in the Bay Area, Santa Clara was down 31%, San Mateo was down 28%, and Alameda was down 25%. We might as well ask why San Francisco was able to buck the trend for as long as it did.

It is well known that women everywhere in America are having fewer children and having them at an older age. San Francisco, as usual, is more extreme than elsewhere.

• 54% of births in 2020-2022 were the mothers’ first births, the highest fraction in the state. The state average is 40%. San Mateo (47%), Santa Clara (47%), and Alameda (46%) were next highest.

• Only 4.3% of births in 2022-2022 were the mothers’ fourth birth or more, the lowest fraction in the state. The state average is 12.1%.

• 83% of births in 2020-2022 were to mothers aged 30 or older, the highest fraction in the state. Marin (81%), San Mateo (76%), Santa Clara (72%) and Alameda (71%) were the only other counties to exceed 70%. The state average was 58%. In Kings County in the Central Valley, it was just under 40%.

• 72% of births in 2020-2022 were to mothers who were college graduates, the highest fraction in the state. Marin (70%), San Mateo (69%), and Santa Clara (62%) were the only other counties to exceed 60%. The state average was 37%. In some Central Valley counties, it’s only 15%.

• The fertility rate, defined as the number of births per woman aged 15-44, is the second-lowest in the state. Only Yolo’s is lower⁸. This is only partially due to the racial composition of the city. Hispanic, White, Asian, and Black women in San Francisco all have fertility rates that are among the lowest in the state. Across California, the fertility rates for those groups were 56, 49, 49, and 45. (I’ll bet few realize Black women have the lowest fertility rates). In San Francisco, the corresponding fertility rates were 50, 43, 37, and 36⁹.

Retention

Births are one of the main drivers of the child population. The other is migration. San Francisco has long been a net exporter of children. Many who give birth in the city move out to the suburbs in search of more space, or a place to buy that they can afford, or schools that they believe to be better. Of course, some children do move in to the city with their families but they are always outnumbered by those who leave. We can get a sense of how big an issue this is by comparing the size of cohorts over time.

We can compare the number of children aged under 5 with the number of births to city residents in the previous five years. This will give a measure of the number who move out of the city even before their kids start school. For example, there were 43,900 births to city residents between April 2015 and March 2020. Those babies would have been 0-4 years old at the time of the census in April 2020 but that census showed only 38,215 kids of that age group living in the city, a net loss of 13%.

We can compare the number aged 5-9 with the number of births to city residents 5-9 years before. For example, the number of births to city residents between April 2010 and March 2015 was 44,723. The 2020 census recorded only 32,594 kids aged between 5 and 9, indicating a drop of 27%. Many of those would have left before the kids started school but there’s no mid-decade census so we don’t know precisely when they left.

We can compare the number aged 10-14 with the number who were in the 0-4 bucket in the previous census. For example, in 2010, there were 35,299 kids in the 0-4 bucket but, ten years later, there were only 30,501 in the 10-14 bucket, indicating a net loss of 14%.

If we look back over time, every cohort has experienced a net loss in every time period. The decline in the most recent decade was actually on the low side. That could be because fewer families left the city or it could be because more immigrants with families arrived. We should hope that it’s the former because that’s more likely to be sustainable.

This pattern makes it a very bad omen for the future that births in San Francisco have fallen so significantly since the pandemic.

Income Levels

For most of the post-war period, San Francisco was not a rich city. The median income of its families¹⁰ tracked the state’s median income. But starting in 2000, San Francisco has become a much richer city. The median family income is now 154% of the state’s median family income.

Battle for Market Share or Increase the Market Size

The biggest factor affecting the number of public school students is not the market share of private schools but the number of school-age children in the city. Any measure that increased the number of school-age children would benefit SFUSD. Are there policies the city could adopt that would encourage families to stay in the city?Remember, we don’t need to persuade families to move to the city. We just need to persuade the families that start here to stay here. Would it help, for example, for the city to demand more two and three bedroom homes and fewer studio and one-bedroom homes? I don’t know which policies would work but it’s unfortunate that this isn’t even a concern for our politicians.

Of course, the perceived quality of the school options available in San Francisco affects a family’s decision about whether to stay in the city. If the product is better, there will be more demand for it. Becoming a better school district is hard. It requires consistent hard work and focus over many years. It requires avoiding chaos and uncertainty and the allure of sociological experiments.

Even if the school district were functioning well, its competitive position would be deteriorating. People whose success is due to their education are the most ardent believers in the importance of education. As we’ve seen, mothers in San Francisco are better educated, are having fewer children, and are having them at an older age, when they are further along in their careers and earning more. Families in San Francisco are richer than ever before. This combination is leading to greater demand for private schools, independent of what the school district does.

1. Only eight cities had populations over 100,000. Today, 76 do.

2. In the state 24.4% of males 14 years old and older, and 15.9% of females, 14 and older, were single. In San Francisco, 29% of males and 19.7% of females were single. In those days, people in rural areas were less likely to be married so San Francisco had the marriage rates of a rural area.

3. Today, there are over 20 cities in California with at least 100,000 people and a higher percentage of children than San Francisco did at its peak back in 1960. The largest of these are Fresno, Bakersfield, and Stockton.

4. If you’re wondering why I didn’t mention this group until now, it’s because 1970 was the first census to count “persons of Spanish language or Spanish surname”. 96% of them identified as White in the racial classification.

5. Intracensal population estimates are going to be less accurate this decade because they are based on estimated changes since the 2020 census and that was conducted during the pandemic and so was less accurate than prior censuses. The census bureau isn’t even using the last census as the base for its estimates but rather a “blended base” that incorporates data not available at the time the census was released. The overall population figure in the blended base is very similar to that in the census but the Under 18 population figure differs by 5%. We have the following estimates of the Under 18 population.

• 2010 Census: 107,524

• 2020 Census: 113,227

• 2020 Blended Base: 118,989

• 2023 Population Estimate: 110,389

Neither the 2020 Census nor the 2020 Blended Base gives credible results.

• The 2020 census figure implies that the U18 population has fallen by less than 3,000 (110,389 - 113,227). We know that’s untrue because public and private school enrollment has fallen by about 5,000 and births are also down.

• The 2020 blended base implies that the U18 population rose during the 2010s by over 11,000 (118,989 - 107,524). If such an increase had happened, it should have shown up in public or private school enrollments. It didn’t.

6. Manhattan has 74,775 people/sq. mile (1,694,251 people in 22.7 square miles) whereas San Francisco has 18,714 people/sql. mile (873,965 people in 46.7 square miles)

7. The most extreme ratios are found in small cities with big prisons such as Florence, AZ (560.1) or Corcoran, CA (337.4). Places with reputations as havens for gay people can also have high ratios e.g. Provincetown, MA (190.5), Palm Springs (182.7), West Hollywood, CA (145.6).

8. Yolo’s low fertility rate may be due to the presence of UC Davis. Colleges contain lots of young women but women of that age are no longer the ones who give birth. Santa Cruz also has a big UC campus and a low fertility rate.

9. San Francisco’s measured fertility rate actually bumped up to 45.9 during the 2020-2022 period. Even though the number of births was down significantly, so many women of childbearing age left the city that the fertility rate increased.

10. Household income also shows a rapid increase since 1990.